CTC and Macnica launch "Physical AI Training": A practical program that provides end-to-end training from digital twin construction to actual robot operation, serving as a substitute for skilled labor.

ITOCHU Techno-Solutions Corporation (Representative Director and President: Tatsushi Shingu, Headquarters: Minato-ku, Tokyo, Abbreviation: CTC) and Macnica (Representative Director and President: Kazumasa Hara, Headquarters: Yokohama City, Kanagawa Prefecture, hereinafter: Macnica) will begin offering "Physical AI Training" utilizing NVIDIA's platform.
This training program utilizes NVIDIA Omniverse (a suite of libraries and development tools that support the construction, simulation, and operation of physical AI applications) to provide a comprehensive package of knowledge and skills necessary for building the foundation of "physical AI" (AI that sees, judges, and acts in the real world). The program includes lectures, practical exercises such as environment setup and robot operation using actual equipment, and aims to deliver 20 orders within three years, primarily targeting manufacturing companies working on the introduction of physical AI.

In Japan, labor shortages and the transfer of technical skills are challenges due to the declining birthrate and aging population. In industries such as manufacturing and logistics, AI that recognizes on-site conditions and acts autonomously is attracting attention as a technology that supports the reproduction and automation of skilled techniques. Unlike conventional AI that learns trends from data, physical AI needs to reproduce skills based on the five senses, such as seeing and touching to make judgments. This requires the construction of a virtual space (digital twin technology) that precisely reproduces physical behavior by integrating sensing that captures the surroundings with high precision, learning algorithms, and feedback from actual equipment.

Physical AI Training is a comprehensive program that teaches essential physical AI skills for robot operation, automation, and autonomous AI, from basic to advanced levels. Leveraging NVIDIA Omniverse and NVIDIA AI Enterprise, a software suite for AI development, you will learn how to develop applications.

This training program begins with learning the fundamentals of physical AI, followed by simulating robot behavior in a digital twin environment using NVIDIA Omniverse. Furthermore, through practical exercises involving motion control and sensor integration using edge devices and actual robots, participants will systematically acquire skills by moving back and forth between the virtual and real worlds. An option to provide a GPU workstation to assist with the start of the training is also available.

This collaboration expands the training provided by CTC in a digital twin environment by adding Macnica 's robot execution environment technology and practical training elements, allowing for consistent learning from virtual to actual robots. In addition, some curriculum elements were jointly developed by both companies, resulting in a program that allows for more practical learning from basic to advanced levels. Both companies are NVIDIA certified partners supporting companies' use of AI, with CTC handling the introduction and operation of Omniverse digital twins, and Macnica handling GPU infrastructure and robot technologies including NVIDIA Isaac.

Moving forward, both companies will continue to expand their curriculum in line with market needs and technological trends, contributing to companies' utilization of physical AI and strengthening their competitiveness as foundational technologies that will support the future.

■Physical AI Training Image Diagram

■Related Sites
　https://www.ctc-g.co.jp/solutions/nvidia/omniverse.html
　https://www.macnica.co.jp/business/semiconductor/manufacturers/nvidia/products/143448/
　https://www.ctc-g.co.jp/solutions/nvidia/ai_enterprise.html
　https://www.macnica.co.jp/business/semiconductor/manufacturers/nvidia/products/143447/

* The product names and other proper nouns mentioned are trademarks or registered trademarks of their respective companies.
* The information provided is current as of the publication date. Please note that it may differ from the latest information.

that's all